Clay Tech

"clay-works make things real"

translated from clazytech.com

Generative AI genuinely saved my KDP release work

Last time, I closed on a sour note: “Editing in Japanese turned out to only be a little easier.” This time, I’ll properly swing the pendulum back. In the practical work of getting the book out on Kindle (KDP), generative AI genuinely saved me.

Same AI—so why did it land so differently? That’s what this post is about.


Actually, this was my second time with KDP

I’ll confess: publishing on KDP wasn’t my first time. I’d done it once before, a long time ago.

But that experience turned out to be completely useless. I’d long since forgotten the procedures, and more importantly, the specs and standards had changed entirely in the meantime. EPUB conventions, paperback submission requirements, the tax paperwork—all of it was a different world from back then. My supposedly “experienced” past knowledge turned out to be almost worthless.

Writing this, I laughed a little at how much it resembled my protagonist, sent back in time with knowledge of the future that doesn’t let him fight as he is. In the end, all you can do is re-access the correct answer for this exact moment, over and over. And that’s exactly where AI showed what it could do.

AI’s “effectiveness” flipped completely between editing and publishing work

What didn’t work in last time’s editing phase was judgment calls like “is this Japanese phrasing acceptable or not,” which ultimately depend on human sensibility. The KDP release work, on the other hand, was an entirely different kind of task. The procedures are fixed, there’s a lot of research involved, and there are specific places you get stuck—in other words, it’s mostly about “the steps and investigation needed to reach the right answer.” This was AI’s home turf.

Where it concretely helped

① Building the EPUB conversion pipeline The manuscript was in Markdown. I converted it to EPUB using Pandoc, and having help hashing out the settings and command structure made a huge difference.

② Troubleshooting validation This phase actually started from a casual question: “Can I check an EPUB using a smartphone app?” The answer came back: smartphones are fine for viewing, but validation is strict, so a PC is safer before publishing. When I followed up with “I’m on a Mac, what should I use?”, it laid out the right tools by purpose.

Settling into a three-step flow—“① spec check → ② device preview → ③ KDP submission”—mattered a lot. Every time an error came up, I got help sorting out what it was and how to fix it, and the time I’d otherwise have spent wandering alone through official documentation disappeared entirely.

③ KDP-related paperwork – Understanding W-8BEN (a US tax document) – Designing the product description (up to 4,000 characters), keywords, and categories – Sorting out KDP Select (exclusive) vs. non-exclusive → Ended up with a hybrid: KDP plus partial free release on Narou and Kakuyomu

④ Cover production workflow I generated a base image with AI image generation (Nano Banana), then layered typography on top with Canva. This included help with building prompts and talking through direction—AI accompanied me through all of it. That said, the cover wasn’t simply a case of “AI helped me.” I wrestled with it quite a bit, and thought seriously about copyright too. I’ll write about that in detail next time.

What actually helped me most was the print book

I’d only planned on an ebook, but greed got the better of me and I made a paperback version too. And this was the area where, left on my own, I would have broken.

This kind of work—plain, technical, the kind where getting stuck eats a whole night—was exactly where running alongside AI paid off by an order of magnitude. I knew what I wanted to do but didn’t know the conventions, so I couldn’t move forward, and AI closed that gap in an instant. It worked in a completely different way from “the quality of the Japanese” last time.

The final boss was “page count equals cost”

Once I’d stamped out all the layout breakage, what stood in my way at the end was print cost. Roughly speaking, KDP paperbacks get more expensive as page count increases. If I could just shave off another dozen-odd pages, the cost would drop a tier, so I entered a final battle fought over a few cents.

The line spacing was already at its limit; tightening it further would ruin the reading experience. That’s when I noticed the “airiness” of web-novel style. That style of breaking into a new paragraph with every line of dialogue or description turns, on the printed page, into nearly 10,000 “gaps between paragraphs,” each one adding whitespace that was quietly devouring pages.

What worked here was, instead of having AI calculate it on paper, repeating “actually typeset it and count” over and over. The result was a bit shocking: without touching the line spacing by even a millimeter, just by tightening the space between paragraphs, the page count dropped by more than 30%. In short, it was simply a matter of moving toward the kind of typesetting Japanese bunko paperbacks have long used—indentation only, no blank lines. Combined with revisiting the trim size, the print book, which originally ran to nearly 550 pages, was ultimately compressed to the 210-page range—less than half.

One quietly significant decision was keeping the ebook (EPUB) and print (PDF) builds separate from the start. Thanks to that, the tightly packed layout I’d built to cut print costs didn’t ruin the reading experience of the ebook at all. The optimal layout changes with the medium. That’s a lesson that sank into my bones after putting out just one book.

So where does AI actually work, and where doesn’t it?

Having written posts three and four now, here’s my own conclusion:

In other words, AI showed what it could really do not in “the story itself” but in “the road to getting the story out into the world.” That’s the realization that clicked into place most for me after putting out one book.


Next time: the story of how that cover came together. I’ll write about the conflicts unique to an AI-generated cover, and the unavoidable copyright considerations.

👉 Read it on Kindle


Originally published in Japanese at https://clazytech.com/2026/06/1648/. Translated with LLM assistance and reviewed before publication.